Research project 1 focuses on identifying response predictors and resistance mechanisms for phosphatidylinositide 3-kinase (PI3K) inhibitors through the conduct of a PDX trial using a broad selection of breast cancer models of hormone receptor positive (HR+)/HER2- and triple negative breast cancer (TNBC) subtypes. PI3K inhibitors are attractive therapeutic agents for breast cancer because of the frequent occurrence of PIK3CA mutations in hormone receptor positive (HR+) and PTEN loss in TNBC. However single agent PI3K inhibitors have yielded only modest anti-tumor activity and the efficacy of pan-PI3K inhibitors are often limited by dose limiting toxicities in clinical trials. Isoform-specific PI3K inhibitors are therefore of great interest to maximize target inhibition with improved tolerability. A major challenge in the development of PI3K inhibitors have been the identification of patients who may benefit the most from these agents and for whom isoform specific inhibitors are appropriate. Clinical trials have focused on limited markers including PIK3CA mutation and PTEN status, which do not consistently predict response. In-depth biomarker research is needed, but is often difficult in the clinical trial setting due to the quantity and quality of the study material and the inability to obtain serial tumor biopsies from patients. In contrast, patient-derived xenograft (PDX) models provide nearly unlimited tumor resources for pre- and post-treatment in-depth genoproteomic analysis for biomarker development and determination of resistance mechanisms for rational drug combinations. We hypothesize that a breast cancer PDX trial of pan- and isoform-specific PI3K inhibitors, will discern isoform dependency of individual tumors to derive predictive biomarkers. We will use copanlisib, a potent pan-PI3K inhibitor with activity against predominantly PI3K? versus PI3K?, and AZD8186, a specific PI3K? inhibitor, both agents are in the NCI-IND portfolio. In addition, based on literature evidence and our preliminary data from a PDX trial of buparlisib, we hypothesize that adding the MEK inhibitor selumetinib, also an NCI IND agent, will improve response to copanlisib and/or AZD8186 based on literature evidence and the synergistic anti-tumor effects we observed when combining a pan-PI3K inhibitor with a MEK inhibitor in our PDX models of breast cancer. Aim 1 will conduct the PDX trial. 100 breast cancer PDX models comprised of ER+HER2- and TNBC subtypes available from the PDX core will be selected based on passage number, genomic stability and tumor characteristics. To enrich for the target population, at least 30 PDX models with PIK3CA mutation and at least 30 PDX models with PTEN null phenotype will be included in this PDX trial. Each PDX model will be passaged to 12 immune-deficient mice and randomly assigned to 6 treatment groups (n=2 each treatment group) to receive either vehicle, copanlisib, AZD8186, selumatinib, copanlisib + selumetinib or AZD8186 + selumetinib. Several tumor response criteria will be used for classification of sensitive vs resistant to drug(s) therapy. These include the recently published mRECIST criteria and the traditional % tumor growth inhibition which includes a vehicle treated group. Aim 2 will identify candidate and novel genoproteomic predictors of response for each single agent or combination therapy by analyzing global genomic data (whole exome and RNA Seq) and proteomic data generated by multi-kinase inhibitor bead mass spectrometry (MIB-MS) for unbiased discovery of candidate biomarkers. The analysis will include a focus on identifying outlying, differentially expressed biomarkers between sensitive and resistant tumors. Aim 3 will perform kinome profiling (MIB-MS) of post treatment samples to assess drug-induced signaling changes in order to discern mechanisms of action of the study treatments and to identify intrinsic and treatment-induced adaptive survival mechanisms. This project will be supported by the infrastructure established by the WU-PDTC for data collection, including drug treatment, mice tolerability and tumor volumes changes over time. In addition, the bioinformatics capability provided by the WU-PDTC for trial interpretation and genoproteomic analysis will be leveraged to derive predictors of response and resistance mechanisms to PI3K inhibitors. In addition, team members of Project 1 will interact regularly with team members of Project 2 for information exchange to improve research approach and results sharing. The long term goal of this research is to set up a standardized PDX trial platform to identify promising drug(s) and biomarker pairs for clinical testing.